考慮工件族整備時間之雙流線型工廠排程

Abstract

本研究探討考慮跨廠與整備時間的雙流線型工廠排程問題。此排程之研究目標為最小化寬裕時間之變異係數，寬裕時間即為交期與完工時間之間的差距。於此問題中，我們將需要相同整備治具的工件集合稱之為工件族。大多數以往的文獻中，不是使用family-based(將全部皆為同ㄧ家族的工件視為單一工件做排程)，就是使用individual-based(每個工件排程是獨立地，而不考慮加入相同工件族)。本研究提出了group-based的方法(亦即將一個工件族區分成多個工件集合，排程中將每個工作集合當作單一個體)。此三個以基因演算法為基礎的GA-EDD-Family、GA-EDD-Group及GA-EDD-Individual方法被發展以及經由多種實驗做比較。實驗結果指出GA-EDD-Group的方法在多數情境下勝過於其他兩種方法。

This research examines a dual flow shop scheduling problem, which is in the context of considering cross-plant processing and setup times. The scheduling objective is to minimize coefficient of variation of slack times, in which the slack time of a job denotes the difference between the due date and its total processing time. Herein, a set of jobs that need the same setup is called a job family. Most prior literature either used a family-based approach (all jobs of a particular family are scheduled as a single job) or used an individual-based approach (each job is independently scheduled without considering its affiliation to its job family). This research proposes a group-based approach (that is, dividing a job family into several job groups, and scheduling each job group as a single entity). Several genetic algorithms (GAs), which are of GA-EDD-Family, GA-EDD-Group, or GA-EDD-Individual, have been developed and compared by numerical experiments. Experiment results indicate that the group-based approach outperforms the other two approaches in most scenarios.